mingk (5) [Avatar] Offline
I own the first edition of the book, but after looking at the second edition table of contents I see a lot of new topics added. Could you expand on the new material you plan to cover?
Anonymous (122) [Avatar]
I guess it's an entirely new book, even by another author and in python scikit instead of Java.
The topics are more or less the same I think, with some new stuff added like deep learning.
Just my 2 cents.
Douglas McIlwraith (12) [Avatar] Offline
Hi mingk,

This is an excellent question and I'm sure many other readers would be keen to know more.

The original book is an excellent text and we wanted to make sure that it remained relevant. With it coming up to 6 years since its release, we felt it deserved a revision.

This is not a simple rewrite, replacing Java for Python, but an extensive update. We still cover the topics of clustering, classification and recommendation but we move away from search. Extra coverage is given to the topic of moving data around the intelligent web as well as to the topics of deep learning and mutli-armed bandits.

I do hope this answers your question but if you have any further queries please feel to post a follow up and I will get back to you.

Happy reading!

milind1944 (7) [Avatar] Offline
Hi Douglas: The first edition had an excellent discussion of naive Bayes classifiers, while this edition omits them altogether. Why is that? I ask because I'm interested in developing an news classifier and was looking forward to a python version in this book. Thanks
ssharma (51) [Avatar] Offline
Douglas McIlwraith (12) [Avatar] Offline

Many thanks for your interest in the book!

The 2nd ed. is quite a different book that the 1st ed. in many respects. We've completely reworked the structure, as you'll see from the Manning site, and included several topics which are not present in the 1st ed., most notably Deep Learning and Multi-Armed Bandits.

Unfortunately, one of the side effects of including new content, is that some of the existing content has been removed. This is not to say that Bayesian classifiers are not equally relevant or important since the publication of the 1st ed., however.

For an accessible introduction (and beyond!) to probability and machine learning I'd recommend looking at Kevin Murphy's textbook: https://www.cs.ubc.ca/~murphyk/MLbook/

Happy reading!
kiquetal (5) [Avatar] Offline
Missing the java code xD
338537 (1) [Avatar] Offline
I have some simple questions and need straightforward answers:

Why you replaced Java with Python?

Any technical reason or it is just a new version for Python developers?

Thanks in advance.